JP2004294953A - Road shape restoring device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a road shape restoring device capable of restoring likely straight lines and likely curves from the limited number of data points by considering the characteristics of map data and that of road structure. <P>SOLUTION: The road shape restoring device comprises a recording medium 10 in which digital shape information is recorded, a data reading means 20 for reading out shape data from the recording medium 10, a 1st interpolated shape calculation means 30 for improving the accuracy of a restored shape by correcting data points while considering the characteristics of map data points especially, and a 2nd interpolated shape calculation means 40 for improving the accuracy of a restored shape especially on the basis of the features of road structure. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

【００３５】
３次Ｂスプライン曲線は、点Ｐ_ｉと点Ｐ_ｉ＋１の間を点Ｐ_ｉ−１〜Ｐ_ｉ＋２の影響をうけるような重み関数を用いて算出される。つまり、前後の点の影響をうけるため、形状変化が大きくかつデータ点数が少ない図５のような状況において、Ｂスプラインで補間した形状は期待される形状からはずれ、元来のデータ点に対しても大きくずれることが考えられる。なお、Ｂスプラインの定義式は式（３）〜（５）で示される。
【数２】

【００３６】
また、図６は３次スプライン曲線を用いてデータ点間を補間した結果を示す。３次スプライン曲線はセグメント間を３次関数で滑らかにつなぐものであり、データ点を必ず通る。しかし、データ点部分において急激な曲線方向の変化を表現することが難しいため、図６のようにデータ点間の補間形状が膨らんだり、大きな誤差を生む可能性がある。
【００３７】
ここで、データ特性判定手段３１１について示す。データ特性判定手段は、期待する形状に対してデータが適当か否かを判定するものである。連続する３点によって得られる隣接２線分の成す角の大きさから補正が必要な状態を判定するための、角度判定手段３１４であってもよいし、連続する３点によって得られる隣接２線分の成す角と、片側の長さから補正が必要な状態を判定するための、データ間隔判定手段３１５であってもよい。あるいは、双方を有し、ＡＮＤ条件で判定をおこなってもよいし、ＯＲで判定を行ってもよい。
【００３８】
また、データ補正算出手段３１２は、データ特性判定手段３１１によって判定された結果に基づき、補正データを算出するものである。角度判定手段３１４の結果に基づき、連続する３点によって得られる２つの線分上に補正点を算出する、第１補正データ算出手段３１６であってもよいし、データ間隔判定手段の結果に基づき、連続する３点によって得られる２つの線分上に補正点を算出する、第２補正データ算出手段３１７であってもよい。あるいは、双方の算出手段を有する補正手段であってもよい。
【００３９】
データ特性判定手段３１１の角度判定手法３１４とデータ補正算出手段３１２の第１補正データ算出手法３１６について、例を用いて説明する。復元手法として３次Ｂスプライン曲線を想定してデータ特性の判定を行う。図７に示すように、線分Ｐ_ｉ−１−Ｐ_ｉと線分Ｐ_ｉ−Ｐ_ｉ＋１による成す角をθ_ｉとする。θ_ｉが規定値θ_ｔｈよりも小さい場合は、補間時に前後の点の影響を大きくうけると考え、補正が必要であると考えられる。
【００４０】
３次Ｂスプライン曲線の場合、補間対象区間の前後１点ずつの影響を受けるため、規定値よりも小さいと判定された場合は、図６に示すように、Ｐ_ｉの代わりに例えば線分Ｐ_ｉ−１−Ｐ_ｉ上に補正点Ｐ’_ｉを配し、線分Ｐ_ｉ−Ｐ_ｉ＋１上に点Ｐ’’_ｉを配する。このようにすることで、急激な形状変化を弱め、点数を増加させることによって前後の点の影響を削減することができる。Ｐ’_ｉとＰ’’_ｉの算出方法は式（６）のような形でもよいし、適切に補正が行われるのならば、特に手法は限定しない。
【数３】

【００４１】
同様に、データ特性判定手段３１１のデータ間隔判定手法３１５とデータ補正算出手段３１２の第２補正データ算出手段３１７について、例を用いて説明する。復元手法は同様に３次Ｂスプライン曲線を想定する。２線分による成す角θ_ｉが規定値より大きい場合であっても、データ点の間隔が広い場合、図５のように補間形状に誤差が生じる可能性がある。そこで、例えば図９に示すように、線分Ｐ_ｉ＋１−Ｐ_ｉ＋２と線分Ｐ_ｉ＋２−Ｐ_ｉ＋３による成す角をθ_ｉ＋２とする。Ｐ_ｉ＋１から線分Ｐ_ｉ＋２−Ｐ_ｉ＋３に平行な線を引いた場合、平行線の間隔をＬ_ｉ＋２とする。Ｌ_ｉ＋２が規定値Ｌ_ｔｈ以上に大きくなると、補間時に前後の点の影響が大きくなると考え、補正が必要であると判定する。ただし、平行線の間隔は式（７）で表現される。
【数４】

【００４２】
同様に規定値よりも大きいと判定された場合は、図１０に示すように、Ｐ_ｉ＋２の代わりに例えば線分Ｐ_ｉ＋１−Ｐ_ｉ＋２上に補正点Ｐ’_ｉ＋２を配し、線分Ｐ_ｉ＋２−Ｐ_ｉ＋３に補正点Ｐ’’_ｉ＋２を配する。Ｐ’_ｉ＋２とＰ’’_ｉ＋２の算出方法は式（８）のような形でもよいし、適切に補正が行われるのならば、特に手法は限定しない。
【数５】

【００４３】
なお、第１補正データ算出手段３１６、第２補正データ算出手段３１７の組み合わせ方は任意とする。元のデータを算出された補正データで更新し、想定した３次Ｂスプライン曲線で補間を行った結果を図１１に示す。図５と比較して、補間形状の誤差が少なくなっていることが分かる。
【００４４】
上記は３次Ｂスプライン曲線を例にとって説明したが、Ｂスプライン曲線に限らず、スプライン曲線、ベジエ曲線、ナーブス曲線でもよく、それぞれの補間手法に適した判定基準や補正方法をとることで、直線部を直線らしく、カーブ部をカーブらしく復元でき、精度のよい形状を復元することが可能になる。判定と補正の手順について図１２の処理フローで示す。
【００４５】
次に、第２補間形状算出手段４０について具体例を用いて説明する。図４は第２の補間形状算出手段４０の構成を示すものであり、道路構造の特徴に基づいて形状を補間するものである。第２の補間形状算出手段４０は、道路構造の特徴を記述した道路形状拘束条件４１と、地図データから得られた形状を道路構造の拘束条件と比較して最適な補間形状を推定する形状推定手段４２からなる。
【００４６】
道路形状拘束条件４１は、例えば道路構造は直線と円弧と緩和曲線の連続で表現され、それぞれの状態の連続や状態間の遷移は一定の条件で制約されたものである。図１３は道路曲率変化が直線と円弧と緩和曲線の連続で表現されている例を示し、図１４はそれぞれの状態が継続する条件と、状態間の遷移が許される条件の例を示すものである。図１３、１４から、道路構造は同じ方向の半径の異なる円弧が連続することはなく、直線部分か、反対方向の円弧を取る。また、曲率変化が一定以下の時以外は、直線部と円弧部、正円弧部と負円弧部の間の遷移には緩和曲線が挿入される。さらに、円弧や緩和曲線はその曲率に応じて、最低連続長が既定されている。道路拘束条件は、上記のような曲率で情報を有してもよいし、道路構造の条件を適切に示す形ならば特にその形態は問わない。
【００４７】
形状推定手段４２は、道路形状拘束条件４１で保持する曲率変化と、地図データから推定される形状の曲率変化とを比較し、拘束条件で有する条件に合うように形状を推定する。たとえば、道路の曲率変化特性に注目して比較する場合、次のような構成が考えられる。地図データから得られる道路の座標データ、あるいは第１の補間形状算出手段３０のデータ補正部３１において補正された補正データから曲率変化を算出するための曲率算出部４３と、算出された曲率変化曲線と道路形状拘束条件から得られる曲率変化特性とを比較し、適切な道路パラメータを推定する道路パラメータ推定部４４と、推定したパラメータから２次元道路形状を算出する形状算出部４５と、算出された二次元道路形状と、地図データから得られる形状とを比較する実形状比較部４６からなる。実形状比較部４６において、推定された形状と地図データから得られる形状とが大きくかけ離れている場合は、道路パラメータ推定部４４において、格差が少なくなるようにパラメータを再度推定する手順を繰り返す。なお、ここでいう道路パラメータとは、曲率の状態（直線、正円弧、負円弧、上り緩和、下り緩和）と、それぞれの状態の開始地点、終了地点、開始位置での曲率、終了位置での曲率を示す。
【００４８】
構成は特にこれに限るわけではなく、道路構造の拘束条件を考慮した形状推定が行われるならば、任意の推定手法を用いてよい。例えば、地図データ点がプロットされた二次元平面上に、制限内の円弧と直線と緩和曲線を当てはめて形状推定を行ってもよい。
【００４９】
図１５は地図データから推定された曲線の曲率変化曲線と、望ましい曲率変化曲線を示すものである。図１５のように、３次スプライン曲線などで補間する場合、滑らかにデータ点間を補間することはできるが、その曲率変化は実際の道路で許容される形状と異なる場合がある。例えば、破線で示す急カーブ→緩カーブ→急カーブのような、一方向の異なる曲率を持つ円弧が連続するような状況は実際の道路では許容されていない。このような部分を、円弧→直線→円弧になるように適切に道路パラメータを推定することで、地図データ点数が不足している状況においても、直線部は直線部として、カーブ部はカーブ部として復元することが可能になる。
【００５０】
図１６に具体的な道路パラメータ推定方法について図解し、推定手順を図１７のフローチャートに示す。推定したい状態は、直線／上り緩和曲線／下り緩和曲線／正円弧／負円弧の５つの状態であり、それぞれの開始地点と曲率変化は線分ｃ＝α_ｉ・ｌ＋β_ｉ（ｌ_ｉ≦ｌ≦ｌ_ｉ＋１）を求めることで推定できる。一定区間ΔＬ（ｌ_ｉ≦ｌ≦ｌ_ｉ＋１）の曲率変化のデータを読み込み、その区間の曲率変化を近似する直線ｃ＝α_ｉ・ｌ＋β_ｉを最小二乗近似などからもとめる。一つの線分の連続区間が一定距離未満の場合は、さらにΔＬ分の区間を読みこんで２ΔＬ区間分の直線近似を行う。連続区間距離が次の曲率状態へ遷移を許容する長さになった場合、現直線を延長し、３ΔＬでの直線と曲線との距離誤差を求める。距離誤差が所定値以下の場合は、さらの近似区間を延長する。距離誤差が一定距離以上となる場合、次の曲率状態への遷移が妥当とし、求めたｃ＝α_ｉ・ｌ＋β_ｉ（ｌ_ｉ≦ｌ≦ｌ_ｉ＋１）から道路の状態、当該状態の開始地点、終点地点、開始地点での曲率、終了地点での曲率を算出する。
【００５１】
次の状態の区間ΔＬ（ｌ_ｉ＋１≦ｌ≦ｌ_ｉ＋２）の曲率変化のデータを読み込み、新に近似直線を求める。一つ前の状態の近似直線と、新たに求められた近似直線との交点を求めることで、線分の開始地点と一つ前の状態の終了地点を再計算する。但し、道路構造拘束条件から、道路の状態は一つ前の道路の状態に依存するため、αとβが取りうる値に制約条件がある。なお、直線部、円弧部は、直線近似の際はα＝０と固定してβを求めればよい。
【００５２】
上記手順を繰り返すことで、道路構造拘束条件を満たすような道路パラメータを求めることができ、求めたパラメータから道路形状を復元することが可能になる。図１７では最急降下法に基づいた近似直線の求め方を示したが、妥当なパラメータを推定できる手法であれば、ニュートン法などのその他の推定手法を用いてもよい。
【００５３】
また、図１８に示すように、道路パラメータ推定と実形状比較部を繰り返すことでより正確に道路形状を復元することが可能となる。
【００５４】
上記のように、ナビゲーション地図データのようなデータ点が限られ、直線部とカーブ部を適切に復元できないような状況においても、上記、第１補間形状算出手段、第２補間形状算出手段、あるいは両方を有する形状復元装置を用いることで、直線部は直線らしく、カーブ部はカーブらしく復元することが可能になる。
【００５５】
尚、本発明の距離分布検知装置は、上述の図示例にのみ限定されるものではなく、本発明の要旨を逸脱しない範囲内において種々変更を加え得ることは勿論である。
【図面の簡単な説明】
【図１】本発明の実施の形態の全体構成を示すブロック図。
【図２】従来技術の問題点を示す図。
【図３】本発明の実施の形態における第１補間形状算出手段の構成を示すブロック図。
【図４】本発明の実施の形態における第２補間形状算出手段の構成を示すブロック図。
【図５】３次Ｂスプラインによる補間で補間誤差が生じることを示す図。
【図６】３次スプラインによる補間で補間誤差が生じることを示す図。
【図７】本発明の実施の形態における角度判定手段の処理を示す図。
【図８】本発明の実施の形態における第１補正データ算出手段の処理を示す図。
【図９】本発明の実施の形態におけるデータ間隔判定手段の処理を示す図。
【図１０】本発明の実施の形態における第２補正データ算出手段の処理を示す図。
【図１１】本発明の実施の形態によって補正を行った結果を示す図。
【図１２】本発明の実施の形態における形状補正の処理の流れを示すフローチャート。
【図１３】道路構造の曲率変化の例を示す図。
【図１４】道路形状拘束条件の例を示す図。
【図１５】地図データから推定された曲線の曲率変化曲線と、望ましい曲率変化曲線を示す図。
【図１６】本発明の実施の形態における道路構造パラメータ推定の処理を示す図。
【図１７】本発明の実施の形態における道路パラメータ推定の処理の流れを示すフローチャート。
【図１８】本発明の実施の形態の道路パラメータ推定と実形状比較との関係を示す図。[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a road shape restoration device for restoring a road shape from map data used for in-vehicle navigation and the like.
[0002]
[Prior art]
Patent Document 1 discloses a vehicle navigation device that can more accurately determine a road curve. In the vehicle navigation device described in Patent Literature 1, a curve is calculated by deriving a curve from four or more points and calculating a radius of curvature of the curve. As a method for calculating a curve, interpolation using a B-spline curve is described.
[0003]
Patent Document 2 discloses a graphic processing apparatus capable of drawing a smooth graphic with less data to be processed. In the graphic processing device described in Patent Literature 2, a road shape is interpolated by a curve based on data points prepared at each scale. As a method for interpolation, an interpolation method using a three-dimensional spline curve is described.
[0004]
[Patent Document 1]
JP 2000-321086 A
[Patent Document 2]
JP 2001-148026 A
[0005]
[Problems to be solved by the invention]
However, although the vehicle navigation device described in Patent Document 1 mentions that the interpolation shape deviates from the data when the road shape is acute, it is intended to correct the location where the curve is reported. Thus, there is no mention of a point at which the interpolation shape is corrected, and the graphic processing device described in Patent Document 2 discloses a point at which the interpolation shape deviates from the data point when an appropriate map coordinate point is not given. Not considered.
[0006]
In general, digital map data installed in a navigation system or the like does not have enough data points to be described in order to reduce data capacity and cost. In that case, when map data is displayed on the navigation screen, it is not smooth if it is connected by a straight line. Therefore, in order to solve this, Patent Literature 1 and Patent Literature 2 describe a method of performing interpolation using a B-spline curve, a cubic spline curve, or the like. However, when the number of data points is limited, one line segment often represents both a straight line portion and a curved portion. Therefore, when performing interpolation using a B-spline curve or the like, if the number of data points representing the shape is insufficient, the interpolated shape may deviate greatly from the expected shape, which is a problem. (See Fig. 2)
[0007]
The present invention has been made to solve the above-described problem, and the object is to consider the characteristics of the map data and the characteristics of the road structure, so that a straight line looks like a straight line from a limited number of data points, It is an object of the present invention to provide a road shape restoration device that can restore a curve like a curve.
[0008]
[Means for Solving the Problems]
The road shape restoring device according to claim 1 of the present invention reads digital shape information having two-dimensional shape information indicated by discrete points, data reading means for reading the shape information, and data read by the data reading means. And a shape interpolation means for calculating an interpolation shape from a point sequence of three or more consecutive data points in accordance with the characteristics of the data points.
[0009]
According to a second aspect of the present invention, there is provided a road shape restoring device, wherein the shape interpolation means calculates an interpolation shape by correcting the data point in consideration of the characteristics of the map data point. And the feature.
[0010]
Further, the road shape restoring device according to claim 3 of the present invention is characterized in that the shape interpolation means has second interpolation shape calculation means for calculating an interpolation shape based on features of the road structure.
[0011]
According to a fourth aspect of the present invention, in the road shape restoring apparatus, the first interpolation shape calculating means calculates the interpolation shape by correcting the data points in consideration of the characteristics of the map data points. And a second interpolation shape calculation means for calculating an interpolation shape based on the features of the road structure and the data points corrected by the first interpolation shape calculation means.
[0012]
Also, in the road shape restoring device according to claim 5 of the present invention, the first interpolation shape calculation means is composed of a data correction means and a data interpolation means, and the data correction means has three or more continuous points. It is characterized by having a data characteristic determining means for determining whether or not correction is necessary from a sequence of data points.
[0013]
According to a sixth aspect of the present invention, there is provided a road shape restoring device, wherein the data correcting means includes a data correction calculating means for correcting and calculating an appropriate data point based on the determination result of the data characteristic determining means. And
[0014]
According to a seventh aspect of the present invention, there is provided the road shape restoring device, wherein the data correction means updates the digital shape data which the data correction means has in advance based on the data corrected and calculated by the data correction calculation means. It is characterized by having means.
[0015]
Further, in the road shape restoring apparatus according to claim 8 of the present invention, the data characteristic determining means determines a state in which correction is necessary based on an angle formed by two adjacent lines obtained by three consecutive data points. It is characterized by having an angle determining means for determining.
[0016]
According to a ninth aspect of the present invention, in the road shape restoring device, the data characteristic determining means includes an angle formed by two adjacent lines obtained by three consecutive data points, and one length of the line segment. And a data interval determination unit for determining a state requiring correction based on the data interval.
[0017]
Further, in the road shape restoring device according to claim 10 of the present invention, the second interpolation shape calculating means uses a road shape constraint condition expressed by a continuation of a straight line, an arc, and a transition curve representing a feature of the road structure. It has a shape estimating means for estimating a road shape.
[0018]
Further, in the road shape restoring device according to claim 11 of the present invention, the shape estimating means indicates the curvature change of the road from the calculated curvature change curve of the road and the calculated curvature change curve of the road. Road parameter estimating means for estimating road parameters, shape calculating means for calculating a road shape from the estimated road parameters, and real shape comparing means for comparing the calculated road shape with a shape obtained from map data It is characterized by.
[0019]
In the road shape restoring device according to claim 12 of the present invention, the shape estimating means determines that the error between the calculated road shape and the shape obtained from the map data is large in the actual shape comparing means. In this case, the road parameter estimating means estimates the road parameter again.
[0020]
Further, another road shape restoring device of the present invention is a first correction data calculating means for calculating correction points one by one on two line segments obtained by three consecutive points based on the result of the angle determining means. It is characterized by having. The other road shape restoring device of the present invention calculates a second correction data for calculating correction points one by one on two line segments obtained by three consecutive points based on the result of the data interval determination means. It is characterized by having means.
[0021]
Function and effect of the present invention
The situation where the interpolated shape from a given data point deviates significantly from the expected shape occurs because the number of data points that determine the shape of the line and curve is not sufficient for the degree of shape change. is there.
[0022]
For example, the angle θ formed by two line segments represented by three points as shown by a portion surrounded by a dashed line in FIG. _i Is an angle close to 90 degrees. These three points (two line segments) are considered to represent a change of a straight line-curve-straight line. However, since the cubic B-spline curve is strongly affected by points before and after the interpolation target section, a steep change cannot be realized in the case of the shape shown in FIG. In the case of interpolation using a cubic spline curve, the tendency of the error is different, but it is also difficult to realize a steep change.
[0023]
Therefore, when it is determined that the data represented by the three points has a shape change of a certain value or more, it is determined that the number of data points is small in order to express the change of the straight line-curve-straight line, and one or more data points are located near the center point. Or add two or more points in place of the center point. By doing so, the number of control points increases, and the influence of the front and rear points is reduced, so that the straight line portion can be restored as a straight line, and the curved portion can be restored as a curve. By determining in advance the case where correction is necessary in accordance with the characteristics of the curve used for interpolation and calculating the correction point, it is possible to improve any correction means.
[0024]
The road is designed based on certain constraint conditions. The road is composed of a straight line and an arc, and a transition curve gently connecting the straight line and the arc. For example, there is a change as shown in FIG. The arc has a road shape constraint condition such that a circular arc continues for a certain section, a straight line does not continue for a certain length or more, and if there is a change in curvature beyond a certain value, a relaxation curve is inserted therebetween. Even if the curve estimated from the map data is smoothly interpolated, the curve may not match the actual characteristics of the road shape change. Therefore, by introducing the road shape constraint condition, it is no longer possible to represent a continuous curve such as a gentle curve-a sharp curve-a gentle curve, and it is possible to restore the straight part as a straight line and the curved part as a curve. .
[0025]
As described above, the road shape restoration device of the present invention assigns data points to both sides near the center when the angle formed by two line segments expressed by three consecutive points is equal to or smaller than a certain value. At this time, the position at which the data point is provided is changed according to the angle formed and the length of the line segment. When the angle formed is small, a data point close to the original data is given, assuming that the radius of curvature is small. Further, the road shape restoration device of the present invention derives a curvature change curve based on the original data and compares it with a road constraint condition. A road model is applied to the obtained curvature change curve to estimate an optimum change pattern. That is, the road shape restoration device of the present invention adds a new point necessary to represent a straight line and a curve according to the interpolation curve, and based on the constraint condition of the road structure, the road is formed by the straight line, the relaxation curve, and the arc. Is a road shape restoring device that can restore a road from a limited number of data points to a straight line as a straight line and a curve as a curved line.
[0026]
BEST MODE FOR CARRYING OUT THE INVENTION
An embodiment of a road shape restoration device according to the present invention will be described with reference to the drawings. FIG. 1 is a block diagram schematically showing components of the present invention. The road shape restoring apparatus according to the present embodiment includes a recording medium 10 on which digital shape information is recorded, data reading means 20 for reading shape data from the recording medium, and data point correction taking into account the characteristics of map data points in particular. The first interpolated shape calculating means 30 is characterized by increasing the accuracy of the restored shape, and the second interpolated shape calculating means 40 is particularly characterized by increasing the accuracy of the restored shape based on the feature of the road structure. Have. In the present embodiment, even in a situation where the number of data points is small, it is possible to restore a straight portion as a straight line and a curved portion as a curved line.
[0027]
The digital shape information 10 is information expressing a road shape by discrete coordinate values, and may be road map data for navigation or a road map for another purpose.
[0028]
The data reading means 20 reads the shape information from the digital shape information 10, and may read the entire data or may sequentially read as much data as necessary.
[0029]
Although the first interpolation shape calculation means 30 and the second interpolation shape calculation means 40 are independently established, the data calculated by the first interpolation shape calculation means 30 may be used as an input to the second interpolation shape calculation means 40.
[0030]
First, the first interpolation shape calculation means 30 will be described. The first interpolation shape calculation means 30 includes a data correction unit 31 that corrects data in consideration of the characteristics of the map data points and a data interpolation unit 32 that interpolates data using a cubic curve or the like based on the corrected data. Become. The data interpolation unit 32 may be a B spline, a spline, a Bezier curve, or the like, or may use another method.
[0031]
FIG. 3 shows the configuration of the data correction unit 31. The data correction unit 31 includes a data characteristic determination unit 311, a data correction unit 312 for correcting data suitable for restoration based on the determination result, and a data update unit that combines the corrected data with the original data. And shape data updating means 313. The data characteristic determining unit 311 determines whether the data is appropriate for the expected shape. The data characteristic determination unit 311 may be common to the interpolation methods assumed by the data interpolation unit 32, or may use a different unit for each interpolation method. The determination result may be expressed by whether or not correction is necessary, or information indicating the degree of correction and the type of correction may be given.
[0032]
The data correction calculator 312 calculates correction data based on the result determined by the data characteristic determiner 311. The correction data calculation may be performed by a common method, or a plurality of methods may be selectively used based on the determination result.
[0033]
The shape data updating means 313 adds and / or replaces the correction data obtained by the data correction calculating means 312 to the original shape data obtained by the data reading means 20, or performs both of them. The shape data updated by the updating unit may be further determined by the data characteristic determining unit 311 and further corrected.
[0034]
Here, an example in which the restored shape deviates greatly from the expected shape will be described. FIG. 5 shows the result of interpolating between given data points using a cubic B-spline curve. The cubic B-spline curve is expressed by the following equations (1) and (2).
(Equation 1)

[0035]
The cubic B-spline curve is represented by the point P _i And point P _{i + 1} Point P between _i-1 ~ P _{i + 2} Is calculated using a weighting function that is influenced by In other words, the shape interpolated by the B-spline deviates from the expected shape in the situation as shown in FIG. 5 where the shape change is large and the number of data points is small because of the influence of the front and rear points. Can also deviate significantly. The definition formula of the B spline is represented by Expressions (3) to (5).
(Equation 2)

[0036]
FIG. 6 shows the result of interpolation between data points using a cubic spline curve. The cubic spline curve connects the segments smoothly with a cubic function and always passes through the data points. However, since it is difficult to express a rapid change in the direction of the curve at the data point portion, there is a possibility that the interpolation shape between the data points expands as shown in FIG. 6 or a large error occurs.
[0037]
Here, the data characteristic determination unit 311 will be described. The data characteristic determining means determines whether the data is appropriate for the expected shape. The angle determination means 314 may be used to determine a state requiring correction based on the magnitude of the angle between two adjacent lines obtained by three consecutive points, or two adjacent lines obtained by three consecutive points. The data interval determination unit 315 may be used to determine a state requiring correction based on the angle formed by the minute and the length of one side. Alternatively, both of them may be used, and the determination may be performed based on the AND condition, or the determination may be performed based on the OR.
[0038]
The data correction calculating means 312 calculates correction data based on the result determined by the data characteristic determining means 311. The first correction data calculation means 316 may calculate correction points on two line segments obtained by three consecutive points based on the result of the angle determination means 314, or may be based on the result of the data interval determination means. The second correction data calculating unit 317 may calculate a correction point on two line segments obtained by three consecutive points. Alternatively, a correction unit having both calculation units may be used.
[0039]
The angle determination method 314 of the data characteristic determination unit 311 and the first correction data calculation method 316 of the data correction calculation unit 312 will be described using examples. The data characteristic is determined assuming a cubic B-spline curve as a restoration method. As shown in FIG. _i-1 −P _i And line segment P _i −P _{i + 1} Angle θ _i And θ _i Is the specified value θ _th If it is smaller than this, it is considered that the influence of points before and after is greatly affected during interpolation, and it is considered that correction is necessary.
[0040]
In the case of a cubic B-spline curve, since it is affected by one point before and after the interpolation target section, if it is determined that it is smaller than the specified value, as shown in FIG. _i Instead of the line segment P _i-1 −P _i Above correction point P ' _i And the line segment P _i −P _{i + 1} Point P '' above _i Distribute. In this way, a sudden change in shape can be weakened, and the number of points can be increased to reduce the influence of the preceding and following points. P ' _i And P '' _i May be calculated as shown in Expression (6), and the method is not particularly limited as long as the correction is appropriately performed.
[Equation 3]

[0041]
Similarly, the data interval determination method 315 of the data characteristic determination unit 311 and the second correction data calculation unit 317 of the data correction calculation unit 312 will be described using examples. The restoration technique also assumes a cubic B-spline curve. Angle θ formed by two line segments _i Is larger than the specified value, there is a possibility that an error occurs in the interpolation shape as shown in FIG. 5 when the interval between the data points is wide. Therefore, for example, as shown in FIG. _{i + 1} −P _{i + 2} And line segment P _{i + 2} −P _{i + 3} Angle θ _{i + 2} And P _{i + 1} From the line segment P _{i + 2} −P _{i + 3} When a line parallel to is drawn, the distance between the parallel lines is L _{i + 2} And L _{i + 2} Is the specified value L _th If it becomes larger than the above, it is considered that the influence of points before and after the interpolation becomes large, and it is determined that correction is necessary. However, the interval between the parallel lines is expressed by equation (7).
(Equation 4)

[0042]
Similarly, when it is determined to be larger than the specified value, as shown in FIG. _{i + 2} Instead of the line segment P _{i + 1} −P _{i + 2} Above correction point P ' _{i + 2} And the line segment P _{i + 2} −P _{i + 3} Correction point P '' _{i + 2} Distribute. P ' _{i + 2} And P '' _{i + 2} May be calculated by the formula (8), and the method is not particularly limited as long as the correction is appropriately performed.
(Equation 5)

[0043]
The combination of the first correction data calculation means 316 and the second correction data calculation means 317 is arbitrary. FIG. 11 shows the result of updating the original data with the calculated correction data and performing interpolation with the assumed cubic B-spline curve. It can be seen that the error of the interpolation shape is reduced as compared with FIG.
[0044]
Although the above description has been made taking the cubic B-spline curve as an example, the present invention is not limited to the B-spline curve, but may be a spline curve, Bezier curve, or Nurbs curve. The portion can be restored as a straight line, and the curved portion can be restored as a curve, and an accurate shape can be restored. The procedure of determination and correction will be described with reference to the processing flow of FIG.
[0045]
Next, the second interpolation shape calculation means 40 will be described using a specific example. FIG. 4 shows the configuration of the second interpolation shape calculation means 40, which interpolates the shape based on the features of the road structure. The second interpolation shape calculating means 40 compares the shape obtained from the map data with the road shape constraint condition 41 describing the feature of the road structure and the shape estimation for estimating the optimal interpolation shape. Means 42.
[0046]
In the road shape constraint condition 41, for example, the road structure is expressed by a continuation of a straight line, an arc, and a relaxation curve, and continuation of each state and transition between the states are restricted under certain conditions. FIG. 13 shows an example in which a change in road curvature is represented by a series of straight lines, circular arcs, and transition curves. FIG. 14 shows an example of conditions in which each state continues and conditions in which transition between states is permitted. is there. As shown in FIGS. 13 and 14, the road structure does not have continuous arcs having different radii in the same direction, and takes a straight line portion or an arc in the opposite direction. Except when the change in curvature is equal to or less than a certain value, a transition curve is inserted in the transition between the straight line portion and the circular arc portion and between the positive circular arc portion and the negative circular arc portion. Further, the minimum continuous length of the arc or the transition curve is predetermined according to the curvature. The road constraint condition may have information with the curvature as described above, and the form is not particularly limited as long as the form appropriately indicates the condition of the road structure.
[0047]
The shape estimating means 42 compares the curvature change held under the road shape constraint condition 41 with the curvature change of the shape estimated from the map data, and estimates the shape so as to meet the condition of the constraint condition. For example, the following configuration is conceivable when making a comparison while paying attention to the curvature change characteristics of the road. A curvature calculation unit 43 for calculating a curvature change from road coordinate data obtained from map data or correction data corrected by the data correction unit 31 of the first interpolation shape calculation unit 30; and a calculated curvature change curve. A road parameter estimating unit 44 for estimating an appropriate road parameter by comparing the characteristic with a curvature change characteristic obtained from the road shape constraint condition, and a shape calculating unit 45 for calculating a two-dimensional road shape from the estimated parameter. An actual shape comparison unit 46 compares the two-dimensional road shape with the shape obtained from the map data. If the estimated shape and the shape obtained from the map data are far apart from each other in the actual shape comparing section 46, the road parameter estimating section 44 repeats the procedure of estimating the parameters again so as to reduce the difference. Note that the road parameters referred to here include the state of curvature (straight line, positive circular arc, negative circular arc, uphill relaxation, downhill relaxation) and the start point, end point, curvature at the start position, and end position of each state. Indicates the curvature.
[0048]
The configuration is not particularly limited to this, and any shape estimation method may be used as long as shape estimation is performed in consideration of the constraint condition of the road structure. For example, shape estimation may be performed by applying an arc, a straight line, and a relaxation curve within the limit on a two-dimensional plane on which map data points are plotted.
[0049]
FIG. 15 shows a curvature change curve of a curve estimated from map data and a desirable curvature change curve. As shown in FIG. 15, when interpolating with a cubic spline curve or the like, it is possible to smoothly interpolate between data points, but the change in curvature may differ from the shape allowed on an actual road. For example, a situation in which arcs having different curvatures in one direction are continuous, such as a sharp curve → a gentle curve → a sharp curve indicated by a broken line, is not allowed on an actual road. By appropriately estimating the road parameters so that such a portion becomes an arc → a straight line → an arc, even in a situation where the number of map data points is insufficient, the straight portion is a straight portion, and the curved portion is a curved portion. It will be possible to restore.
[0050]
FIG. 16 illustrates a specific road parameter estimation method, and the estimation procedure is shown in a flowchart of FIG. The states to be estimated are five states: straight line / upward relaxation curve / downward relaxation curve / positive circular arc / negative circular arc. The starting point and the curvature change of each state are line segment c = α _i ・ L + β _i (L _i ≦ l ≦ l _{i + 1} ) Can be estimated. The fixed section ΔL (l _i ≦ l ≦ l _{i + 1} ) Is read, and a straight line c = α approximating the curvature change in that section. _i ・ L + β _i From the least squares approximation. If the continuous section of one line segment is shorter than the fixed distance, a section of ΔL is read, and a straight line approximation of 2ΔL section is performed. When the continuous section distance has a length that allows the transition to the next curvature state, the current straight line is extended, and the distance error between the straight line and the curve at 3ΔL is obtained. If the distance error is equal to or less than the predetermined value, the approximate section is extended. If the distance error is equal to or greater than a certain distance, the transition to the next curvature state is appropriate, and the obtained c = α _i ・ L + β _i (L _i ≦ l ≦ l _{i + 1} ), The state of the road, the start point, the end point, the curvature at the start point, and the curvature at the end point of the state are calculated.
[0051]
The next state section ΔL (l _{i + 1} ≦ l ≦ l _{i + 2} ) Is read, and a new approximate straight line is obtained. By finding the intersection of the approximate line of the previous state and the newly obtained approximate line, the start point of the line segment and the end point of the previous state are recalculated. However, since the state of the road depends on the state of the immediately preceding road from the road structure restriction conditions, there are restrictions on the possible values of α and β. It should be noted that, for a straight line portion and an arc portion, β may be obtained by fixing α = 0 at the time of straight line approximation.
[0052]
By repeating the above procedure, road parameters that satisfy the road structure constraint condition can be obtained, and the road shape can be restored from the obtained parameters. FIG. 17 shows a method of obtaining an approximate straight line based on the steepest descent method, but any other estimating method such as the Newton method may be used as long as a proper parameter can be estimated.
[0053]
In addition, as shown in FIG. 18, it is possible to more accurately restore the road shape by repeating the road parameter estimation and the actual shape comparison unit.
[0054]
As described above, even in a situation where data points such as navigation map data are limited and a straight line portion and a curved portion cannot be properly restored, the first interpolation shape calculation means, the second interpolation shape calculation means, or By using the shape restoring device having both, it is possible to restore the straight portion as a straight line and the curved portion as a curve.
[0055]
It should be noted that the distance distribution detecting device of the present invention is not limited to the illustrated example described above, and it is needless to say that various changes can be made without departing from the gist of the present invention.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an overall configuration of an embodiment of the present invention.
FIG. 2 is a diagram showing a problem of the related art.
FIG. 3 is a block diagram showing a configuration of a first interpolation shape calculation unit according to the embodiment of the present invention.
FIG. 4 is a block diagram showing a configuration of a second interpolation shape calculation unit according to the embodiment of the present invention.
FIG. 5 is a diagram showing that an interpolation error occurs in interpolation using a cubic B-spline.
FIG. 6 is a diagram showing that an interpolation error occurs in interpolation using a cubic spline.
FIG. 7 is a diagram showing processing of an angle determination unit according to the embodiment of the present invention.
FIG. 8 is a diagram showing processing of a first correction data calculating unit according to the embodiment of the present invention.
FIG. 9 is a diagram showing processing of a data interval determination unit according to the embodiment of the present invention.
FIG. 10 is a diagram showing a process of a second correction data calculating unit according to the embodiment of the present invention.
FIG. 11 is a diagram showing a result of performing correction according to the embodiment of the present invention.
FIG. 12 is a flowchart illustrating a flow of a shape correction process according to the embodiment of the present invention.
FIG. 13 is a diagram showing an example of a change in curvature of a road structure.
FIG. 14 is a diagram showing an example of a road shape constraint condition.
FIG. 15 is a diagram showing a curvature change curve of a curve estimated from map data and a desirable curvature change curve.
FIG. 16 is a diagram showing a process of estimating a road structure parameter in the embodiment of the present invention.
FIG. 17 is a flowchart illustrating a flow of processing for estimating a road parameter according to the embodiment of the present invention.
FIG. 18 is a diagram showing a relationship between road parameter estimation and actual shape comparison according to the embodiment of the present invention.

Claims (12)

Digital shape information having two-dimensional shape information indicated by discrete points;
Data reading means for reading the shape information,
A road shape restoration device, comprising: a shape interpolation unit that calculates an interpolation shape according to the characteristics of the data points from a point sequence of three or more consecutive data points read by the data reading unit. 2. The road shape according to claim 1, wherein said shape interpolation means has a first interpolation shape calculation means for calculating an interpolation shape by correcting a data point in consideration of characteristics of a map data point. Restoration device. 2. The road shape restoration device according to claim 1, wherein the shape interpolation unit includes a second interpolation shape calculation unit that calculates an interpolation shape based on a feature of the road structure. 3. A first interpolation shape calculation means for calculating an interpolation shape by correcting a data point in consideration of characteristics of a map data point;
2. The road according to claim 1, further comprising a data point corrected by the first interpolation shape calculation means and a second interpolation shape calculation means for calculating an interpolation shape based on a feature of the road structure. Shape restoration device. The first interpolation shape calculation means includes a data correction means and a data interpolation means,
5. The road shape according to claim 2, wherein the data correction unit includes a data characteristic determination unit that determines whether correction is necessary based on a sequence of three or more consecutive data points. 6. Restoration device. 6. The road shape restoration device according to claim 5, wherein the data correction unit includes a data correction calculation unit that corrects and calculates an appropriate data point based on a determination result of the data characteristic determination unit. 7. The road according to claim 6, wherein said data correction means has shape data updating means for updating digital shape data which is provided in advance, based on data corrected and calculated by said data correction calculation means. Shape restoration device. The data characteristic determining means includes an angle determining means for determining a state requiring correction from a magnitude of an angle formed by two adjacent segments obtained by three consecutive data points. The road shape restoration device according to any one of claims 5 to 7. The data characteristic determining means determines a data interval required for correction based on an angle formed by two adjacent line segments obtained by three consecutive data points and a length of one of the line segments. The road shape restoring device according to any one of claims 5 to 8, further comprising means. The second interpolation shape calculating means includes a shape estimating means for estimating a road shape from a road shape constraint condition expressed by a series of straight lines, arcs, and relaxation curves representing features of the road structure. Item 4. A road shape restoration device according to item 3. The shape estimating means, a curvature calculating means for calculating a curvature change curve of the road,
Road parameter estimation means for estimating a road parameter indicating a curvature change of the road from the calculated curvature change curve of the road,
Shape calculation means for calculating a road shape from the estimated road parameters,
The road shape restoration device according to claim 10, further comprising actual shape comparison means for comparing the calculated road shape with a shape obtained from map data. The shape estimating means estimates the road parameters again in the road parameter estimating means when the actual shape comparing means determines that the error between the calculated road shape and the shape obtained from the map data is large. The road shape restoration device according to claim 11, wherein the road shape is restored.

Images